Table 1 Holistic innovation comparison across research dimensions.
Research dimension | Prior approach limitations (selected examples) | DIS-HARM innovation |
|---|---|---|
Dynamic Adaptation | •Static carbon coins ([1]) •Fixed pricing ([2]) •Non-adaptive MARL ([3]) | Meta-learning reward coefficients​ (Eq. 5): Real-time cₚ/cₒ updates |
User Heterogeneity Modeling | • Linear VOT ([9]) • Discrete grouping ([10]) • Rigid probit models ([15]) | ​ηᵢ elasticity + Sigmoid βᵢ​ (Eq. 7,11): Nonlinear cross-modal response |
Network Synergy | • Static maturity models ([18]) • TF-IDF recommendations ([20]) • Centrality-based assessment ([21]) | Weather-modulated propagation​ (φ(wₜ) in Eq. 13): Dynamic influence decay |
Multi-objective Optimization | • Efficiency-only focus ([3, 20]) • Unbalanced subsidies ([12]) | ​Fairness-efficiency trade-off​ |
